An Economy Running on Fewer Legs

The U.S. economy right now feels a bit like a bar stool that has lost a few of its legs. As long as the remaining ones hold, everything looks stable enough. But the fewer supports you have, the more vulnerable the whole thing becomes. That is increasingly how the current economic expansion looks. The headlines still show growth, unemployment is relatively low, and consumer spending has not collapsed. Yet beneath those numbers, the economy appears to be relying on a surprisingly small number of engines.

Some of the confusion in recent data is technical. A long federal government shutdown late last year disrupted several data releases, making short term comparisons messy. Seasonal adjustments and year over year comparisons will likely look odd for a few months while the data settles back into a normal rhythm. Even in the best of times, economists caution against drawing conclusions from a single monthly report; even more so in the current environment.

But even when we look past the noise, the structure of the economy looks unusual. Growth is happening, but it is concentrated. In many areas that traditionally power expansions, momentum is surprisingly weak. One of the most striking examples is the labor market. Outside of healthcare, job growth has been extremely limited over the past year. In fact, most of the hiring that has taken place has come from healthcare and related services, particularly roles that support an aging population such as home health aides and personal care workers. These jobs are important and necessary, but they do not usually drive broad economic expansions on their own.

At the same time, total hours worked across the economy have barely grown over the past year. Real disposable income, once government transfers are removed, has also been mostly flat. Those are not numbers that typically describe a booming economy.

Yet the economy continues to expand, the main reason being a massive wave of investment in artificial intelligence infrastructure. Spending on data centers, chips, and AI related computing has surged at a pace rarely seen outside of major technological shifts. The scale of the investment is enormous. In terms of its direct contribution to economic growth, the current AI buildout may be closer to the railroad expansion of the nineteenth century than to more recent technology cycles like the early internet boom.

This spending shows up directly in GDP through construction activity, equipment purchases, and corporate capital expenditures. It also has an indirect effect through financial markets. The same technology companies leading the AI investment wave have helped drive strong stock market performance over the past several years.

That matters because wealth drives spending in the United States. When stock prices rise, households that own financial assets tend to feel more comfortable spending. Since wealth ownership is heavily concentrated among higher income households, much of the spending strength we see today reflects that group’s financial gains.

At the other end of the income spectrum, the picture looks very different (ed. note: this is the "K-shaped" economy you may have heard about). Many lower income households have seen their savings cushions shrink dramatically. The national savings rate has fallen to roughly 3.5 percent, a level that suggests many families are simply trying to maintain their current standard of living rather than building financial buffers. Credit card usage has increased, and consumer sentiment surveys show particularly low confidence among lower income groups.

This creates a strange dynamic in the economy. Spending remains solid in aggregate, but the reasons behind it are uneven. Wealthy households are spending because their balance sheets are strong, while many lower income households are spending because they feel they have little choice.

That divergence helps explain why economic sentiment looks so weak even while the economy continues to grow. Surveys consistently show consumers feeling uneasy about the future. Businesses report caution about hiring and investment. Yet retail spending has held up better than many analysts expected.

Part of the confusion also comes from the data itself. Labor market statistics have become harder to interpret in recent years. Survey response rates have fallen, and the companies most likely to respond tend to be larger and more stable employers. Smaller firms, which are often responsible for a large share of job creation and destruction, are less consistently represented.

Recent benchmark revisions illustrate the challenge. Initial data suggested roughly one million jobs were created in 2025. Updated estimates indicate the number may have been closer to two hundred thousand. That kind of revision does not mean anyone intentionally misreported the data. It simply reflects how difficult it is to measure a complex and constantly changing economy in real time.

From a high level, the economy still appears stable. Unemployment is around 4.3 percent, which historically remains a relatively low level. Labor force participation is decent. None of the classic recession signals are flashing red. But if you dig a little deeper, the trend lines raise some questions. Unemployment has drifted higher over the past year. Hiring has slowed. Wage growth is still positive, but it is not accelerating in a way that would strongly support household spending.

Normally, a combination of weak job growth, low savings, and cautious sentiment would raise stronger recession concerns. What makes the current moment different is the presence of unusually powerful counterweights. The wealth generated by financial markets and the scale of AI investment are both supporting the economy in ways that were not present in past cycles.

The big question is how long those supports can continue.

AI investment, for example, has been growing at an extraordinary pace. But its contribution to economic growth depends on continued expansion. If spending on AI infrastructure simply stabilizes rather than increases further, its incremental contribution to GDP drops to zero. To keep boosting growth, the investment levels would need to continue rising significantly each year.

That is possible, but it becomes harder the larger the spending already is. Even the largest technology companies face limits on how quickly they can expand capital spending, especially if financial markets begin to question the long term payoff. This is why expectations matter so much in the current environment. Corporate earnings have been extremely strong, particularly among technology companies. Profit margins remain historically high, and those profits continue to support equity valuations.

As long as investors believe those earnings trends will continue, financial markets can remain supportive of economic growth. But expectations can change quickly, especially in industries driven by new technology. For now, the outlook remains balanced. The economy is still growing. Inflation has eased from its peak. Financial markets remain supportive. But the underlying structure of the expansion is unusual enough to deserve attention. The economy is not collapsing, but it is leaning heavily on a few key supports. Artificial intelligence investment, strong corporate profits, and wealth driven spending are doing a lot of the work that broader job growth would normally do.